sphere_max_based_geometric_compression

Bases: sphere_patch_based_geometric_compression

The sphere patch and max metric based geometric compression function.

It performs the data compression based on sphere geometric patch shapes and provided max metric. This class inherits from the sphere_patch_based_geometric_compression class, and only redefines the initialization method.

...

Attributes:

Name	Type	Description
name	str, default = 'sphere_max_based_geometric_compression'	Name of the compression function.

Methods:

Name	Description
__init__	It performs the initialization of the sphere compression function based on the max metric.

Source code in tinybig/compression/geometric_compression.py

class sphere_max_based_geometric_compression(sphere_patch_based_geometric_compression):
    """
        The sphere patch and max metric based geometric compression function.

        It performs the data compression based on sphere geometric patch shapes and provided max metric.
        This class inherits from the sphere_patch_based_geometric_compression class, and only redefines the initialization method.

        ...

        Attributes
        ----------
        name: str, default = 'sphere_max_based_geometric_compression'
            Name of the compression function.

        Methods
        ----------
        __init__
            It performs the initialization of the sphere compression function based on the max metric.
    """
    def __init__(self, name: str = 'sphere_max_based_geometric_compression', metric: Callable[[torch.Tensor], torch.Tensor] = None, *args, **kwargs):
        """
            The initialization method of the sphere patch and max metric based geometric compression function.

            It initializes the sphere compression function with the max compression metric.

            Parameters
            ----------
            name: str, default = 'sphere_max_based_geometric_compression'
                Name of the sphere mean compression function.

            Returns
            ----------
            transformation
                The sphere mean geometric compression function.
        """
        super().__init__(name=name, metric=batch_max, *args, **kwargs)

__init__(name='sphere_max_based_geometric_compression', metric=None, *args, **kwargs)

The initialization method of the sphere patch and max metric based geometric compression function.

It initializes the sphere compression function with the max compression metric.

Parameters:

Name	Type	Description	Default
name	str	Name of the sphere mean compression function.	'sphere_max_based_geometric_compression'

Returns:

Type	Description
transformation	The sphere mean geometric compression function.

Source code in tinybig/compression/geometric_compression.py

def __init__(self, name: str = 'sphere_max_based_geometric_compression', metric: Callable[[torch.Tensor], torch.Tensor] = None, *args, **kwargs):
    """
        The initialization method of the sphere patch and max metric based geometric compression function.

        It initializes the sphere compression function with the max compression metric.

        Parameters
        ----------
        name: str, default = 'sphere_max_based_geometric_compression'
            Name of the sphere mean compression function.

        Returns
        ----------
        transformation
            The sphere mean geometric compression function.
    """
    super().__init__(name=name, metric=batch_max, *args, **kwargs)